Disclosed is an improved method, system, and computer program product for implementing an author profiling tool for receiving data created by a plurality of authors, profiling the plurality of authors by performing semantic analysis upon the data, generating a plurality of author profiles for identifying topics of interest to the author based upon results from the semantic analysis and correlating topics of interest by analyzing the plurality of author profiles to identify common topics between the plurality of authors. These author profiles can be used to identify and correlate topical interests by consumers. An enterprise or business can more effectively market to the consumers based upon this knowledge of the consumers' interests.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
1. A computer implemented method for analyzing data comprising: receiving, at an author profiling tool, data created by a group of authors; performing a multi-stage profiling for the group of authors for determining a targeted presentation to at least some of the group of authors, the multi-stage profiling comprising: from within the group of authors, performing a first stage profiling to determine a first subgroup from the group of authors at least by: determining a first topic of interest from the data created by the group of authors at least by: performing a themes analysis upon the data created by the group of authors to create multiple themes and respective strengths for the multiple themes; and determining the first topic based at least in part upon the multiple themes and the respective strengths from the themes; and determining the first subgroup having first multiple authors from the group of authors based at least in part upon respective affinities of the first multiple authors for the first topic of interest; from within the first subgroup, performing a second stage profiling to determining a second subgroup from the first subgroup at least by: determining a second topic of interest at least by performing a correlation analysis on the data created by the first multiple authors in the first subgroup based at least in part upon the multiple themes; and determining the second subgroup having second multiple authors from within the first subgroup based at least in part upon respective degrees of interest of the second multiple authors for the second topic of interest, wherein the second topic of interest is common to the second multiple authors in the second subgroup, and the second topic is different from the first topic; and presenting, to the first subgroup having the first multiple authors, the targeted presentation pertaining to the second topic of interest that is common to the second subgroup.
This invention relates to a computer-implemented method for analyzing data to profile authors and deliver targeted content. The method addresses the challenge of efficiently identifying and engaging specific groups of authors based on their interests within a larger dataset. The system receives data created by a group of authors and performs a multi-stage profiling process to determine targeted presentations for subsets of the group. In the first stage, the system analyzes the data to identify themes and their strengths, then determines a primary topic of interest based on these themes. It then identifies a first subgroup of authors who have a strong affinity for this topic. In the second stage, the system performs a correlation analysis on the data from the first subgroup to identify a secondary topic of interest, distinct from the primary topic, that is common among a smaller subset of authors. The system then presents targeted content related to this secondary topic to the first subgroup, ensuring relevance to the identified interests. This approach enables precise targeting of content to authors with shared but nuanced interests, improving engagement and personalization.
2. The method of claim 1 , further comprising generating actionable data using at least some of a plurality of author profiles, wherein the first topic of interest is determined without targeting specific groups of authors and without referencing information about the first topic of interest, wherein generating the plurality of author profiles further comprises: identifying a theme within the data that is created by the group of authors; and generating a respective strength number corresponding to an extent or degree of interest by the author for the first topic of interest, wherein the data is received at the author profiling tool and is not targeting specific demographic groups, the first topic of interest is determined from the data without targeting the specific demographic groups, and the first topic of interest is determined with the themes analysis without preconceived notions about the data in the themes analysis.
This invention relates to a method for generating actionable data from author profiles without targeting specific demographic groups or preconceived notions. The method involves analyzing data created by a group of authors to identify themes within the content, where the data is collected without demographic targeting. From this analysis, a first topic of interest is determined based on the identified themes, without referencing prior information about the topic or the authors. The method further generates a strength number for each author, representing their degree of interest in the first topic. This strength number quantifies the extent to which an author engages with or contributes to discussions around the topic. The process ensures that the topic identification is unbiased, relying solely on the themes derived from the data rather than external assumptions or demographic filters. The resulting actionable data can be used for various applications, such as content recommendation, trend analysis, or audience engagement strategies, while maintaining an objective and untargeted approach.
3. The method of claim 1 , further comprising: correlating the first subgroup with the second subgroup in response to an identification of the second topic of interest, wherein the data is not targeting specific demographic groups of authors; classifying the data created by the group of authors into a plurality of classes based in part or in whole upon topics of interest determined by the themes analysis, classifying the data including: creating a set of themes from results of the themes analysis; determining multiple subjects of multiple topics of interest based in part or in whole upon the set of themes; determining similarity among the multiple subjects of the multiple topics of interest at least by analyzing the plurality of author profiles; clustering the multiple topics of interest into the plurality of classes based in part or in whole upon the similarity among the multiple subjects; determining the respective strengths for the group of authors, a respective strength for an author of the group of authors indicating relative affinity of the author to a category in the data relative to one or more remaining categories in the data; associating the respective strengths that respectively correspond to the group of authors with a plurality of categories; and creating a first vector for each author of the first subgroup of authors, wherein vectors for the group of authors respectively indicate the respective affinities among the first multiple authors to one or more common topics of interests or one or more subjects.
This invention relates to analyzing and classifying data generated by groups of authors based on topics of interest and author affinities. The method involves correlating two subgroups of authors in response to identifying a second topic of interest, where the data is not targeted at specific demographic groups. The data is classified into multiple classes by first creating a set of themes from a themes analysis. Multiple subjects of interest are then determined based on these themes, and similarity among these subjects is analyzed by examining author profiles. The topics of interest are clustered into classes based on this similarity. The method also determines the relative strengths of authors within the group, indicating their affinity to specific categories in the data compared to other categories. These strengths are associated with multiple categories, and a vector is created for each author in the first subgroup, representing their affinities to common topics or subjects. This approach enables structured analysis of author-generated content without demographic targeting, facilitating insights into thematic relationships and author engagement.
4. The method of claim 3 , further comprising: modifying the plurality of classes determined from classifying the data created by the group of authors at least by reducing a false positive, a false negative, and inappropriate content with a filtering process; identifying actionable data based in part or in whole upon a result of the filtering process, wherein the data created by the group of authors includes contents transcribed from non-social data; determining, at a rule and workflow module stored at least partially in memory, a plurality of computing systems to receive the actionable data based in part or in whole upon a set of rules that identifies how the actionable data is to be processed and directed; preconfiguring a plurality of types of topics of interest; determining a first set of authors that corresponds to one or more first types of topics of interest of the plurality of types of topics of interest at least by analyzing the plurality of author profiles to identify a first set of author profiles respectively corresponding to the first set of authors; determining second commonality within one or more second types of topics of interest of the plurality of types of topics of interest without pre-defining the one or more second types of topics of interest; identifying first commonality among the data in response to the one or more second types of topics of interest based in part or in whole upon results of the themes analysis; identifying a group of authors that corresponds to a first affinity for a first subject; determining a second affinity and a third affinity shared by at least a threshold percentage of authors of the group of authors at least by analyzing a subset of author profiles corresponding to the group of authors and at least by performing one or more first correlation analyses, wherein the second affinity and the third affinity are unknown or unexpected prior to determining the second and the third affinities; generating correlation data based in part or in whole upon a result of determining the second affinity and the third affinity; and generating an action for the group of authors based at least in part on the second affinity and the third affinity, wherein an author profile of the plurality of author profiles comprises a vector comprising a value for the first topic of interest and indicating an affinity or a strength pertaining to the first topic between the author and a different author of the group of authors.
This invention relates to a system for analyzing and processing data created by a group of authors, particularly from non-social data sources such as transcribed content. The system classifies the data into multiple classes and applies a filtering process to reduce false positives, false negatives, and inappropriate content. Actionable data is identified based on the filtering results and directed to specific computing systems according to predefined rules and workflows. The system also preconfigures various topics of interest and analyzes author profiles to determine authors associated with these topics. Additionally, it identifies commonalities and affinities among authors, including unexpected or unknown affinities, by performing correlation analyses. The system generates correlation data and actions based on these affinities. Author profiles are represented as vectors, where each vector includes values indicating an author's affinity or strength of connection to a topic or other authors. The invention aims to improve data processing accuracy and uncover meaningful patterns in author behavior and content.
5. The method of claim 4 , further comprising: performing semantic filtering on the data for reducing irrelevant data from the data, wherein the themes analysis comprises a latent semantic analysis that analyzes contextual and semantic significance of one or more terms that appear within the data; identifying the set of themes from the data based in part or in whole upon a result of the themes analysis and a result of classifying the data; generating or updating the plurality of author profiles for the data based in part or in whole upon the respective strengths for the group of authors and further based at least in part upon the set of themes; identifying a set of rules from a rulebase accessible by the author profiling tool; dispatching, at a rule and workflow engine, actionable data for the group of authors to the plurality of computing systems based in part or in whole upon the set of rules, wherein a rule provides how the actionable data is to be dispatched; determining contextual and semantic significance in the data created by the group of authors at least by performing classification and filtering on the data; and identifying one or more specific themes within the data based in part or in whole upon one or more topics and one or more subjects revealed from the themes analysis and the classification.
This invention relates to a method for analyzing and processing data generated by a group of authors to identify themes, classify content, and dispatch actionable insights. The method addresses the challenge of extracting meaningful information from large datasets by reducing irrelevant data through semantic filtering. It employs latent semantic analysis to assess the contextual and semantic significance of terms within the data, helping to identify relevant themes. The data is classified, and these classifications, along with the semantic analysis results, are used to generate or update author profiles. These profiles are based on the strengths of the authors and the identified themes. The method also involves retrieving a set of rules from a rulebase and using a rule and workflow engine to dispatch actionable data to computing systems based on these rules. The rules dictate how the actionable data is distributed. Additionally, the method determines the contextual and semantic significance of the data by performing classification and filtering, and it identifies specific themes by analyzing topics and subjects revealed through the semantic analysis and classification processes. This approach enhances data processing efficiency and ensures that relevant insights are effectively communicated to the appropriate systems.
6. The method of claim 5 , further comprising: categorizing the multiple topics and the multiple subjects of the data to create a number of categories; associating the respective strengths with the number of categories, the respective strength indicating relative affinities of the first subgroup of authors to a particular topic, a particular subject, or a particular theme; and defining or updating the vector for each of the authors using at least the respective strengths and the number of categories, the vector establishing an author profile for a specific author and being used to describe and analyze the specific author with respect to one or more affinities of the specific author.
This invention relates to analyzing and categorizing authors based on their affinities to topics, subjects, or themes within a dataset. The method involves processing data to identify multiple topics and subjects, then categorizing them into distinct groups. Each author is associated with strengths that indicate their relative affinities to these categories. These strengths are used to define or update a vector for each author, which serves as an author profile. The vector quantifies the author's affinities and enables analysis of their engagement with specific topics, subjects, or themes. The method may also involve clustering authors into subgroups based on their affinities, allowing for deeper insights into author behavior and preferences. The vector-based profiles facilitate comparisons between authors and help identify patterns in their contributions to the dataset. This approach is useful in fields like content analysis, social network analysis, and recommendation systems, where understanding author affinities is critical. The invention improves upon prior methods by providing a structured, quantifiable way to represent and analyze author profiles.
7. The method of claim 6 , wherein the determining the second topic of interest further comprises: identifying first multiple author profiles respectively corresponding to the first multiple authors based at least in part on the first topic of interest, without referencing first information about the first topic of interest; identifying the second topic of interest at least by performing the correlation analysis that analyzes second multiple author profiles, without referencing second information or about the second topic of interest, wherein the second topic of interest is common among the second multiple authors corresponding to the second multiple author profiles; identifying one or more authors in response to an identification of the second topic of interest based in part or in whole upon one or more vectors or one or more relative strengths of the one or more authors with respect to the second topic of interest; and establishing an author-to-author correlation between the one or more authors and the specific author in response to the identification of the second topic of interest.
This invention relates to a method for identifying and correlating topics of interest among authors in a digital content analysis system. The problem addressed is the need to discover relationships between authors based on shared interests without relying on direct information about those topics. The method involves analyzing author profiles to determine a second topic of interest that is common among multiple authors. First, multiple author profiles are identified based on a previously determined first topic of interest, but without using explicit information about that topic. Next, a correlation analysis is performed on a second set of author profiles to identify a second topic of interest shared among those authors, again without referencing direct information about the second topic. The analysis may involve evaluating vectors or relative strengths of authors with respect to the second topic. Once the second topic is identified, one or more authors are selected based on their association with it. Finally, an author-to-author correlation is established between these authors and a specific author, linking them through the shared topic of interest. This approach enables the discovery of indirect relationships between authors based on common interests, improving content recommendation, social network analysis, or targeted communication systems.
8. A computer program product embodied on a non-transitory computer usable medium having stored thereon a sequence of instructions which, when executed by a processor, causes the processor to execute a set of acts for analyzing data, the set of acts comprising: receiving, at an author profiling tool, data created by a group of authors; determining a first topic of interest from the data created by the group of authors at least by: profiling the group of authors at least by performing themes analysis upon the data for generating a plurality of author profiles; generating the plurality of author profiles for identifying the first topic of interest based at least in part upon a result from the themes analysis; and determining the first topic at least by analyzing the plurality of author profiles; determining a first subgroup having first multiple authors in the group of authors at least by classifying the first multiple authors into the first subgroup based at least in part upon the first topic; for the first subgroup, determining a second topic of interest common to second multiple authors in the first subgroup at least by performing a correlation analysis that analyzes at least some profiles of the first multiple authors in the first subgroup; and for the second topic of interest, determining a second subgroup having second multiple authors from the first subgroup, wherein the second topic of interest is common to the second subgroup of authors, and the second topic is different from the first topic.
This invention relates to a computer-implemented method for analyzing data created by a group of authors to identify topics of interest and subgroup authors based on those topics. The system receives data from a group of authors and performs theme analysis to generate author profiles, which are used to identify a first topic of interest. The authors are then classified into a first subgroup based on their association with this topic. Within this subgroup, a correlation analysis is performed on the author profiles to identify a second, distinct topic of interest. A second subgroup is then formed from authors within the first subgroup who share this second topic. The method enables hierarchical topic-based clustering of authors, allowing for deeper analysis of specialized interests within broader groups. The system leverages profiling and correlation techniques to refine topic identification and subgroup formation iteratively. This approach is useful for applications such as content recommendation, targeted communication, or audience segmentation where understanding nested interest structures is valuable. The invention is implemented as a computer program product with instructions stored on a non-transitory medium, executed by a processor to perform the described analysis steps.
9. The computer program product of claim 8 , further comprising generating actionable data using at least some of the plurality of author profiles, wherein the first topic of interest is determined without targeting specific groups of authors and without preconceived notions about the data in the themes analysis, wherein generating the plurality of author profiles further comprises: identifying a theme within the data that is created by the group of authors; and generating a respective strength number corresponding to an extent or degree of interest by the author for the first topic of interest, wherein the data is received at the author profiling tool and is not targeting specific demographic groups, the first topic of interest is determined from the data without targeting the specific demographic groups, and the first topic of interest is determined with the themes analysis without preconceived notions about the data in the themes analysis.
This invention relates to a computer program product for analyzing author-generated data to identify topics of interest without demographic targeting or preconceived biases. The system receives data from a group of authors and processes it through an author profiling tool to generate actionable insights. The tool identifies themes within the data that reflect collective interests, then creates individual author profiles by calculating a strength number for each author, representing their degree of interest in a particular topic. The topic of interest is determined organically from the data, without pre-selecting specific demographic groups or imposing predefined assumptions during analysis. This approach ensures unbiased discovery of emerging trends by analyzing natural patterns in the data rather than filtering or segmenting authors based on external criteria. The system avoids targeting specific groups, allowing for broader, more inclusive insights into collective interests. The strength numbers quantify each author's engagement with the topic, enabling further analysis or decision-making based on the derived profiles. The method ensures that the identified topics and author interests are derived purely from the data, free from external biases or demographic constraints.
10. The computer program product of claim 8 , the set of acts further comprising: correlating the first subgroup with the second subgroup in response to an identification of the second topic of interest, wherein the data is not targeting specific demographic groups of authors; classifying the data created by the group of authors into a plurality of classes based in part or in whole upon topics of interest determined by the themes analysis, classifying the data including: creating a set of themes from results of the themes analysis; determining multiple subjects of multiple topics of interest based in part or in whole upon the set of themes; determining similarity among the multiple subjects of the multiple topics of interest at least by analyzing the plurality of author profiles; clustering the multiple topics of interest into the plurality of classes based in part or in whole upon the similarity among the multiple subjects; determining the respective strengths for the group of authors, a respective strength for an author of the first subgroup of authors indicating relative affinities of the first subgroup to a category in the data relative to one or more remaining categories in the data; associating the respective strengths that respectively correspond to the group of authors with a plurality of categories; and creating a first vector for each author of the group of authors, wherein vectors for the group of authors respectively indicate respective affinities among the first subgroup to one or more common topics of interests or one or more subjects.
This invention relates to analyzing and classifying data created by groups of authors based on topics of interest and author profiles, without targeting specific demographic groups. The system performs a themes analysis to identify topics of interest and generates a set of themes from the results. Multiple subjects related to these topics are determined, and their similarity is assessed by analyzing author profiles. The topics are then clustered into multiple classes based on this similarity. The system also determines the relative strengths of authors within subgroups, indicating their affinities to different categories in the data. These strengths are associated with various categories, and a vector is created for each author to represent their affinities to common topics or subjects. The first subgroup of authors is correlated with a second subgroup in response to identifying a second topic of interest. This approach enables the classification of data into meaningful categories while avoiding demographic targeting, facilitating better understanding of author interests and contributions.
11. The computer program product of claim 10 , the set of acts further comprising: modifying the plurality of classes determined from classifying the data created by the group of authors at least by reducing a false positive, a false negative, and inappropriate content with a filtering process; identifying actionable data based in part or in whole upon a result of the filtering process, wherein the data created by the group of authors includes contents transcribed from non-social data; determining, at a rule and workflow module stored at least partially in memory, a plurality of computing systems to receive the actionable data based in part or in whole upon a set of rules that identifies how the actionable data is to be processed and directed; preconfiguring a plurality of types of topics of interest; determining a first set of authors that corresponds to one or more first types of topics of interest of the plurality of types of topics of interest at least by analyzing the plurality of author profiles to identify a first set of author profiles respectively corresponding to the first set of authors; determining second commonality within one or more second types of topics of interest of the plurality of types of topics of interest without pre-defining the one or more second types of topics of interest; identifying first commonality among the data in response to the one or more second types of topics of interest based in part or in whole upon results of the themes analysis; identifying a group of authors that corresponds to a first affinity for a first subject; determining a second affinity and a third affinity shared by at least a threshold percentage of authors of the group of authors at least by analyzing a subset of author profiles corresponding to the group of authors and at least by performing one or more first correlation analyses, wherein the second affinity and the third affinity are unknown or unexpected prior to determining the second and the third affinities; generating correlation data based in part or in whole upon a result of determining the second affinity and the third affinity; and generating an action for the group of authors based at least in part on the second affinity and the third affinity, wherein an author profile of the plurality of author profiles comprises a vector comprising a value for the first topic of interest for the author and indicating an affinity or a strength pertaining to the first topic between the author and a different author of the group of authors.
This invention relates to a system for analyzing and processing data created by a group of authors, particularly from non-social data sources, to identify actionable insights and direct them to appropriate computing systems. The system improves data classification by reducing false positives, false negatives, and inappropriate content through a filtering process. It identifies actionable data based on the filtered results and determines which computing systems should receive this data using predefined rules and workflows. The system preconfigures various topics of interest and analyzes author profiles to identify authors associated with specific topics. It also detects commonalities within topics without prior definitions, analyzing themes to uncover shared interests. Additionally, the system identifies groups of authors with affinities for particular subjects and performs correlation analyses to discover unexpected or unknown affinities among authors. These affinities are quantified and used to generate actionable insights for the group. Author profiles are represented as vectors, containing values that indicate an author's affinity or strength of connection to a topic or other authors. The system leverages these profiles to enhance data analysis and decision-making processes. The overall goal is to improve the accuracy and relevance of data processing by dynamically identifying patterns and directing insights to the appropriate systems for further action.
12. The computer program product of claim 11 , further comprising: performing semantic filtering on the data for reducing irrelevant data from the data, wherein the themes analysis comprises a classification that classifies the data into multiple themes and a latent semantic analysis that analyzes contextual and semantic significance of one or more terms that appear within the data; identifying the set of themes from the data based in part or in whole upon a result of the themes analysis and a result of classifying the data; generating or updating the plurality of author profiles for the data based in part or in whole upon the respective strengths for the group of authors and further based at least in part upon the set of themes; identifying a set of rules from a rulebase accessible by the author profiling tool; dispatching, at a rule and workflow engine, actionable data for the group of authors to the plurality of computing systems based in part or in whole upon the set of rules, wherein a rule provides how the actionable data is to be dispatched; determining contextual and semantic significance in the data created by the group of authors at least by performing classification and filtering on the data; and identifying one or more specific themes within the data based in part or in whole upon one or more topics and one or more subjects revealed from the themes analysis and the classification.
This invention relates to a computer program product for analyzing and processing data generated by a group of authors to improve content relevance and workflow efficiency. The system performs semantic filtering to reduce irrelevant data by classifying the data into multiple themes and conducting latent semantic analysis to assess the contextual and semantic significance of terms within the data. Themes are identified based on the results of this analysis and classification. Author profiles are generated or updated based on the strengths of the authors and the identified themes. The system accesses a rulebase to identify a set of rules, which dictate how actionable data is dispatched to computing systems via a rule and workflow engine. The system also determines the contextual and semantic significance of the data by performing classification and filtering, and identifies specific themes by analyzing topics and subjects revealed through the themes analysis and classification. This approach enhances data relevance, automates workflows, and improves content management by leveraging semantic and thematic insights.
13. The computer program product of claim 12 , the set of acts further comprising: categorizing the multiple topics and the multiple subjects of the data to create a number of categories; associating the respective strengths with the number of categories, the respective strength indicating relative affinity of each author of the first subgroup to a particular topic, a particular subject, or a particular theme; and defining or updating the vector for each author using at least the respective strengths and the number of categories, the vector establishing an author profile for a specific author and being used to describe and analyze the specific author with respect to one or more affinities of the specific author.
This invention relates to analyzing and categorizing authors based on their affinities to topics, subjects, or themes within a dataset. The problem addressed is the need to quantify and profile authors' interests or expertise in a structured way, enabling better understanding of their contributions and relevance to specific areas. The system processes data containing multiple topics and subjects, categorizing them into distinct groups. Each author in a subgroup is assigned a strength value indicating their relative affinity to a particular topic, subject, or theme. These strengths are then used to define or update a vector representing the author's profile. The vector serves as a quantitative measure of the author's affinities, allowing for analysis of their interests or expertise in a structured format. The categorization and strength association steps ensure that the author profiles are comprehensive and reflect nuanced relationships between authors and the content they engage with. The vector-based representation enables efficient comparison and clustering of authors based on their affinities, supporting applications in recommendation systems, expertise mapping, or content personalization. The invention improves upon prior methods by providing a dynamic, multi-dimensional profile that evolves with new data.
14. The computer program product of claim 13 , the set of acts further comprising: determining the second topic of interest common to the second subgroup having the second multiple authors, without targeting specific demographic groups of authors or referencing preconceived information about the data for identifying the second topic of interest, at least by: identifying first multiple author profiles corresponding to the first multiple authors based at least in part on the first topic of interest, without preconceived notions about the data for identifying the first topic of interest; identifying the second topic of interest at least by performing the correlation analysis that analyzes second multiple author profiles in the first multiple author profiles without the preconceived information about the data for identifying the second topic of interest, wherein the second topic of interest is common to second multiple authors corresponding to the second multiple author profiles; identifying one or more authors in response to an identification of the second topic of interest based in part or in whole upon one or more vectors or one or more relative strengths of the one or more authors with respect to the second topic of interest; and establishing a correlation between the one or more authors and the specific author in response to the identification of the second topic of interest.
This invention relates to a method for identifying topics of interest within a group of authors, particularly in a way that avoids preconceived biases or demographic targeting. The system first identifies a first topic of interest common to a first subgroup of authors by analyzing their profiles without relying on preconceived notions about the data. It then performs a correlation analysis on a subset of these profiles to determine a second topic of interest that is also common to a second subgroup of authors within the first group. This second topic is identified without targeting specific demographic groups or referencing preexisting assumptions about the data. The system further identifies individual authors associated with the second topic based on their relative strengths or vectors related to that topic. Finally, it establishes a correlation between these identified authors and a specific author of interest, all while maintaining an unbiased approach to topic discovery. The method ensures that topic identification remains objective and data-driven, avoiding external biases that could skew results.
15. A computer system for analyzing data, comprising: a computer processor to execute a set of program code instructions; and a memory to hold the set of program code instructions, in which the set of program code instructions comprises instructions for: receiving, at an author profiling tool, data created by a group of authors; determining a first topic of interest from the data created by the group of authors at least by: profiling the group of authors at least by performing themes analysis upon the data for generating a plurality of author profiles; generating the plurality of author profiles for identifying the first topic of interest based at least in part upon a result from the themes analysis; and determining the first topic at least by analyzing the plurality of author profiles; determining a first subgroup having first multiple authors in the group of authors at least by classifying the first multiple authors into the first subgroup based at least in part upon the first topic; for the first subgroup, determining a second topic of interest common to second multiple authors in the first subgroup at least by performing a correlation analysis that analyzes at least some profiles of the first multiple authors in the first subgroup; and for the second topic of interest, determining a second subgroup having second multiple authors from the first subgroup, wherein the second topic of interest is common to the second subgroup of authors, and the second topic is different from the first topic.
This invention relates to a computer system for analyzing data created by a group of authors to identify topics of interest and subgroup authors based on those topics. The system uses a computer processor and memory to execute program code instructions that perform several key functions. First, it receives data created by a group of authors and processes this data through an author profiling tool. The tool performs themes analysis on the data to generate multiple author profiles, which are then used to identify a first topic of interest. The system determines this topic by analyzing the generated profiles. Next, the system classifies authors into a first subgroup based on their association with the first topic. For this subgroup, it performs a correlation analysis on the profiles of the authors to identify a second topic of interest that is common to a subset of authors within the first subgroup. The system then determines a second subgroup from the first subgroup, where the second topic is different from the first. This approach allows for hierarchical topic-based clustering of authors, enabling deeper insights into shared interests within nested subgroups. The system leverages computational analysis to automate the identification of topics and author groupings, improving efficiency in data-driven author profiling.
16. The computer system of claim 15 , the set of program code instructions further comprising instructions for generating actionable data using at least some of the plurality of author profiles, wherein the first topic of interest is determined without targeting specific demographic groups of users and without preconceived notions about the data for identifying the first topic of interest, wherein the set of program code instructions for generating the plurality of author profiles further comprise the instructions for: identifying a theme within the data that is created by the group of authors; and generating a respective strength number corresponding to an extent or degree of interest by the author for the first topic of interest, wherein the data is received at the author profiling tool and is not targeting the specific demographic groups, the first topic of interest is determined from the data with the themes analysis without targeting the specific demographic groups, and the first topic of interest is determined from the data with the themes analysis without preconceived notions about the data in the themes analysis.
This invention relates to a computer system for analyzing author-generated data to identify emerging topics of interest without demographic targeting or preconceived biases. The system processes data from a group of authors to generate author profiles, which include themes extracted from the data and corresponding strength numbers indicating the degree of interest in a particular topic. The system determines a first topic of interest by analyzing themes within the data, where the analysis is performed without pre-selecting demographic groups or imposing preconceived assumptions about the data. The author profiling tool receives unsolicited data, meaning the data is not collected with specific demographic targeting in mind. The system then generates actionable insights based on the identified topics and author profiles, enabling unbiased discovery of trends and interests within the data. The approach ensures that the identified topics are organically derived from the data itself, rather than being influenced by external factors or predefined categories. This method is particularly useful for uncovering emerging trends in social media, forums, or other user-generated content platforms where understanding natural interest patterns is valuable.
17. The computer system of claim 15 , wherein the set of program code instructions for generating the plurality of author profiles further comprises the instructions for: correlating the first subgroup with the second subgroup in response to an identification of the second topic of interest, wherein the data is received from the group of authors without targeting specific demographic groups of authors; classifying the data created by the group of authors into a plurality of classes based in part or in whole upon topics of interest determined by the themes analysis, classifying the data including: creating a set of themes from results of the themes analysis; determining multiple subjects of multiple topics of interest based in part or in whole upon the set of themes; determining similarity among the multiple subjects of the multiple topics of interest at least by analyzing the plurality of author profiles; clustering the multiple topics of interest into the plurality of classes based in part or in whole upon the similarity among the multiple subjects; determining the respective strengths for the group of authors, a respective strength for an author of the group of authors indicating a relative affinity of the author to a category in the data relative to one or more remaining categories in the data; associating the respective strengths that respectively correspond to the group of authors with a plurality of categories; and creating a first vector for each author of the group of authors, wherein vectors for the group of authors respectively indicate respective affinities among the first subgroup to one or more common topics of interests or one or more subjects.
This invention relates to a computer system for analyzing and classifying data generated by a group of authors, particularly for identifying and correlating topics of interest without targeting specific demographic groups. The system generates author profiles by correlating subgroups of data based on identified topics, classifying the data into multiple classes, and determining the relative strengths or affinities of authors to different categories within the data. The classification process involves creating themes from themes analysis, determining multiple subjects and topics of interest, and clustering these topics into classes based on similarity among the subjects. The system also calculates the respective strengths of authors, indicating their relative affinity to specific categories compared to others, and associates these strengths with various categories. Additionally, the system creates vectors for each author, representing their affinities to common topics or subjects, enabling further analysis of author interests and data patterns. The approach allows for the automated grouping and categorization of author-generated content based on thematic analysis and affinity measurements, facilitating insights into collective interests and trends.
18. The computer system of claim 17 , wherein the set of program code instructions further comprises instructions for: modifying the plurality of classes determined from classifying the data created by the group of authors at least by reducing a false positive, a false negative, and inappropriate content with a filtering process; identifying actionable data based in part or in whole upon a result of the filtering process, wherein the data created by the group of authors includes contents transcribed from non-social data; determining, at a rule and workflow module stored at least partially in memory, a plurality of computing systems to receive the actionable data based in part or in whole upon a set of rules that identifies how the actionable data is to be processed and directed; preconfiguring a plurality of types of topics of interest; determining a first set of authors that corresponds to one or more first types of topics of interest of the plurality of types of topics of interest at least by analyzing the plurality of author profiles to identify a first set of author profiles respectively corresponding to the first set of authors; determining second commonality within one or more second types of topics of interest of the plurality of types of topics of interest without pre-defining the one or more second types of topics of interest; identifying first commonality among the data in response to the one or more second types of topics of interest based in part or in whole upon results of the themes analysis; identifying a group of authors that corresponds to a first affinity for a first subject; determining a second affinity and a third affinity shared by at least a threshold percentage of authors of the group of authors at least by analyzing a subset of author profiles corresponding to the group of authors and at least by performing one or more first correlation analyses, wherein the second affinity and the third affinity are unknown or unexpected prior to determining the second and the third affinities; generating correlation data based in part or in whole upon a result of determining the second affinity and the third affinity; and generating an action for the group of authors based at least in part on the second affinity and the third affinity, wherein an author profile of the plurality of author profiles comprises a vector comprising a value for the first topic of interest and indicating an affinity or a strength pertaining to the first topic between the author and a different author of the group of authors.
A computer system analyzes data created by a group of authors to identify actionable insights. The system classifies the data, which may include non-social content, and applies a filtering process to reduce false positives, false negatives, and inappropriate content. Actionable data is identified based on this filtering, and a rule and workflow module determines which computing systems should receive and process this data according to predefined rules. The system preconfigures topics of interest and identifies authors associated with these topics by analyzing author profiles. It also detects commonalities within other topics without prior definition, using theme analysis to identify shared content patterns. The system further identifies groups of authors with affinities for specific subjects, analyzing their profiles to uncover unexpected or unknown affinities, such as a second and third affinity shared by a threshold percentage of authors. Correlation data is generated from these affinities, and actions are recommended for the group based on the identified relationships. Author profiles are represented as vectors, where each vector includes values indicating an author's affinity or strength of connection to a topic or another author. This structured approach enables the system to derive meaningful insights from large datasets, improving decision-making and content moderation.
19. The computer system of claim 18 , the set of program code instructions further comprising instructions for: performing semantic filtering on the data for reducing irrelevant data from the data, wherein the themes analysis comprises a classification that classifies the data into multiple themes and a latent semantic analysis that analyzes contextual and semantic significance of one or more terms that appear within the data; identifying the set of themes from the data based in part or in whole upon a result of the themes analysis and a result of classifying the data; generating or updating the plurality of author profiles for the data based in part or in whole upon the respective strengths for the group of authors and further based at least in part upon the set of themes; identifying a set of rules from a rulebase accessible by the author profiling tool; dispatching, at a rule and workflow engine, actionable data for the group of authors to the plurality of computing systems based in part or in whole upon the set of rules, wherein a rule provides how the actionable data is to be dispatched; determining contextual and semantic significance in the data created by the group of authors at least by performing classification and filtering on the data; and identifying one or more specific themes within the data based in part or in whole upon one or more topics and one or more subjects revealed from the themes analysis and the classification.
A computer system analyzes data to identify themes and generate author profiles, improving data relevance and workflow automation. The system performs semantic filtering to reduce irrelevant data by classifying content into multiple themes and conducting latent semantic analysis to assess the contextual and semantic significance of terms within the data. It identifies themes based on this analysis and classification, then generates or updates author profiles by evaluating the strengths of authors and associating them with the identified themes. The system accesses a rulebase to retrieve rules that dictate how actionable data should be dispatched to computing systems. A rule and workflow engine processes these rules to distribute the data accordingly. Additionally, the system determines the contextual and semantic significance of data created by authors through classification and filtering, further refining theme identification by analyzing topics and subjects revealed in the analysis. This approach enhances data organization, author profiling, and automated workflow management.
20. The computer system of claim 15 , the set of program code instructions further comprising instructions for: determining the second topic of interest common to the second subgroup having the second multiple authors, without targeting specific demographic groups of authors or preconceived information about the data for identifying the second topic of interest, at least by: identifying first multiple author profiles corresponding to the first multiple authors based at least in part on the first topic of interest, without preconceived notions about the data for identifying the first topic of interest; identifying the second topic of interest at least by performing the correlation analysis that analyzes second multiple author profiles in the first multiple author profiles without the preconceived information about the data for identifying the second topic of interest, wherein the second topic of interest is common to second multiple authors corresponding to the second multiple author profiles; identifying one or more authors in response to an identification of the second topic of interest based in part or in whole upon one or more vectors or one or more relative strengths of the one or more authors with respect to the second topic of interest; and establishing a correlation between the one or more authors and a specific author in response to the identification of the second topic of interest.
This invention relates to a computer system for analyzing author profiles to identify topics of interest within groups of authors, without relying on preconceived demographic information or assumptions about the data. The system determines a second topic of interest common to a subgroup of authors by first identifying author profiles associated with a previously identified first topic of interest. The system then performs a correlation analysis on these profiles to uncover the second topic, which is shared by a subset of the authors. The analysis does not target specific demographic groups or use preconceived notions about the data. Once the second topic is identified, the system evaluates authors based on their association with this topic, using vectors or relative strengths to quantify their relevance. The system then establishes a correlation between these authors and a specific author, linking them through the shared topic. This approach enables the discovery of latent connections between authors and topics without prior biases, improving the accuracy of topic-based author analysis.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
March 27, 2017
March 8, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.